(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Demonstration of Communication using Neutrinos"

February 26, 2013 20:58 WSPC/INSTRUCTION FILE 
Neutrino' to 'MPL A 



Modern Physics Letters A 

© World Scientific Publishing Company 



DEMONSTRATION OF COMMUNICATION USING NEUTRINOS 



D. D. Standi 1 , P. Adamson 2 , M. Alania 3 , L. Aliaga 4 , M. Andrews 2 , C. Araujo Del Castillo 4 , 
L. Bagby 2 , J.L. Bazo Alba 4 , A. Bodek 5 , D. Boehnlein 2 , R. Bradford 5 , W.K. Brooks 6 , H. Budd 5 , 

A. Butkevich 7 , D.A.M. Caicedo 8 , D.P. Capista 2 , CM. Castromonte 8 , A. Chamorro 3 , 

E. Charlton 9 , M.E. Christy 10 , J. Chvojka 5 , P.D. Conrow 5 , I. Danko 11 , M. Day 5 , J. Devan 9 , 

J.M. Downey 12 , S.A. Dytman 11 , B. Eberly 11 , J.R. Fein 11 , J. Felix 13 , L. Fields 14 , 
G.A. Fiorentini 8 , A.M. Gago 4 , H. Gallagher 15 , R. Gran 16 , J. Grange 17 , J. Griffin 5 , T, Griffin 2 , 
E. Hahn 2 , D.A. Harris 2 , A. Higuera 13 , J.A. Hobbs 14 , CM. Hoffman 5 , B.L. Hughes 1 , 
K. Hurtado 3 , A. Judd 5 , T. Kafka 15 , K. Kephart 2 , J. Kilmer 2 , M. Kordosky 9 , S.A. Kulagin 7 , 
V.A. Kuznetsov 14 , M. Lanari 16 , T. Le 18 , H. Lee 5 , L. Loiacono 5 - 19 , G. Maggi 6 , E. Maher 20 , 
S.Manly 5 , W.A. Mann 15 , CM. Marshall 5 , K.S. McFarland 5 ' 2 , A. Mislivec 5 , A. M. McGowan 5 , 
J.G. Morfm 2 , H. da Motta 8 , J. Mousseau 17 , J.K. Nelson 9 , J.A. Niemiec-Gielata 5 , N. Ochoa 4 , 

B. Osmanov 17 , J. Osta 2 , J.L. Palomino 8 , J.S. Paradis 5 , V. Paolone 11 , J. Park 5 , C. Pena 6 , 
G. Perdue 5 , C.E. Perez Lara 4 , A.M. Peterman 14 , A. Pla-Dalmau 2 , B. Pollock 9 , F. Prokoshin 6 , 
R.D. Ransome 18 , H. Ray 17 , M. Reyhan 18 , P. Rubinov 2 , D. Ruggiero 5 , O.S. Sands 12 , 
H. Schellman 14 , D.W. Schmitz 2 , E.C Schulte 18 , C. Simon 21 , C.J. Solano Salinas 3 , R. Stefanski 2 , 
R.G. Stevens 19 , N. Tagg 22 , V. Takhistov 18 , B.G. Tice 18 , R.N. Tilden 14 , J.P. Velasquez 4 , 
I. Vergalosova 18 , J. Voirin 2 , J. Walding 9 , B.J. Walker 14 , T. Walton 10 , J. Wolcott 5 , 
T.P. Wytock 14 , G. Zavala 13 , D. Zhang 9 , L.Y. Zhu 10 , and B.P. Ziemer 21 

1 Department of Electrical and Computer Engineering, North Carolina State University, Raleigh, 

NC 27695, 

2 Fermi National Accelerator Laboratory, Batavia, IL 60510, 
3 Universidad Nacional de Ingenieria, Av. Tupac Amaru 210, Lima, Peru, 
^Seccion Fisica, Departamento de Ciencias, Pontificia Universidad Catolica del Peru, Apartado 

1761, Lima, Peru, 

5 Department of Physics and Astronomy, University of Rochester, Rochester NY 14627. 
6 Departamento de Fisica, Universidad Tecnica Federico Santa Maria, Avda. Espana 1680 

Casilla 110-V Valparaiso, Chile, 
7 Institute for Nuclear Research of the Russian Academy of Sciences, 117312 Moscow, Russia, 
8 Centro Brasileiro de Pesquisas Fisicas, Rua Dr. Xavier Sigaud 150, Urea, Rio de Janeiro, RJ, 

22290-180, Brazil, 

9 Department of Physics, College of William & Mary, Williamsburg, VA 23187, 
10 Hampton University, Dept. of Physics, Hampton, VA 23668, 
11 Department of Physics and Astronomy, University of Pittsburgh, Pittsburgh, PA 15260, 
12 NASA Glenn Research Center, Cleveland OH 44135, 
13 Departamento de Fisica, Universidad de Guanajuato, Campus Leon, Lomas del Bosque 103, 
fracc. Lomas del Campestre Leon GTO. 37150, Mexico, 
14 Northwestern University, Evanston, IL 60208, 
15 Physics Department, Tufts University, Medford, MA 02155, 
16 Department of Physics, University of Minnesota - Duluth, Duluth, MN 55812, 

17 University of Florida, Department of Physics, Gainesville, FL 32611, 

18 Rutgers, The State University of New Jersey, Piscataway, NJ 08854, 

19 Department of Physics, University of Texas, 1 University Station, Austin, TX 78712. 

20 Massachusetts College of Liberal Arts, 375 Church St., North Adams, MA 01247, 
21 Department of Physics and Astronomy, University of California, Irvine, Irvine, CA 

92697-4575, 

22 Otterbein College, One Otterbein College, Westerville, OH, 43081 



February 26, 2013 20:58 WSPC/INSTRUCTION FILE Neutrino'to'MPLA 



ddstancil@ncsu. edu 

Beams of neutrinos have been proposed as a vehicle for communications under unusual 
circumstances, such as direct point-to-point global communication, communication with 
submarines, secure communications and interstellar communication. We report on the 
performance of a low-rate communications link established using the NuMI beam line 
and the MINERvA detector at Fermilab. The link achieved a decoded data rate of 0.1 
bits/sec with a bit error rate of 1% over a distance of 1.035 km, including 240 m of earth. 

Keywords: Neutrino; Communication. 

PACS Nos.: 14.60.Lm, 01.20.-fx 

The use of fundamental particles that interact very weakly with matter has 
been proposed as a vehicle for communication in environments where electromag- 
netic waves are dampe d and do n ot penetrate easily. The bulk of these proposals 
have involved neutrinosP^E3^E3 although recent consideration has also been given 
to possible use of hypothetical particles such as the axio rJ3 and the hidden-sector 
photonP Neutrino communication systems have been proposed for point-to-point 
global communication s g lobal communication with submarines and for inter- 
stellar communication! 1 ^ Neutrino communication systems can also be contem- 
plated for use in planetary exploration during periods in which the communications 
link is blocked by a planetary body. Even low bandwidth global point-to-point 
links could be useful for secure exchange of encryption codes. While the ability 
to penetrate matter is an important advantage of neutrinos, the weak interactions 
of neutrinos also imply that very intense beams and massive detectors would be 
required to realize this type of communication. 

We report on the performance of a low-rate communications link established us- 
ing the NuMI beam line and the MINERvA detector at Fermilab. The link achieved 
a decoded data rate of 0.1 bits/sec with a bit error rate of 1% over a distance of 
1.035 km that included 240 m of earth. This demonstration illustrates the feasibil- 
ity of using neutrino beams to provide a low-rate communications link, independent 
of any existing electromagnetic communications infrastructure. However, given the 
limited range, low data rate, and extreme technologies required to achieve this goal, 
significant improvements in neutrino beams and detectors are required for "practi- 
cal" application. 

The Neutrino beam at the Main Injector at Fermilab is currently one 

of the most intense high energy neutrino beams worldwide, and provides a variable- 
energy beam for use in particle physics experiments. A simplified schematic of the 
beam line is shown in Fig.[T] A series of accelerators currently produces 8.1 /zs pulses 
of 120 GeV protons every 2.2 s. The repetition rate is limited by the time required to 
accelerate the protons to high energy. The proton beam strikes a carbon target that 
produces many pions, kaons, and other particles. Charged particles are focused using 
magnets to produce a beam that is directed toward detectors. Almost all pions and 
kaons decay into neutrinos (mostly muon neutrinos and other associated particles) 
in a 675 m helium-filled decay pipe, in roughly the same direction as the momentum 



2 



26,2013 20:58 WSPC/INSTRUCTION FILE Neutrino'to'MPLA 



DEMONSTRATION OF COMMUNICATION USING NEUTRINOS 3 



of the original meson. This beam of assorted elementary particles passes through 240 
meters of rock (mostly shale), and all particles are absorbed, except the neutrinos 
that define the NuMI beam. Since proton-carbon interactions at 120 GeV produce 
pions much more readily than kaons, the beam is predominantly composed of muon 
neutrinos (88%) with smaller components of muon antineutrinos (11%) and electron 
neutrinos (1%). The energy spectrum peaks at 3.2 GeV, and has a width (FWHM) 
of about 2.8 GeV. Although the energies extend to 80 GeV, 91.6% of the neutrinos 
in the beam have energies below 10 GeV. The detector is slightly more than 1 km 
from the target, and, at its peak energy, the beam has a transverse size of a few 
meters at the face of the detector. 




1035 m 



Fig. 1. Layout of the NuMI beam line used as the neutrino source, and the MINERvA detector. 



The MINERvA detector (Fig. 12) is located in a cavern about 100 m underground. 
Cross sections for neutrino interactions are measured by observing the trajectories 
of particles emitted when neutrinos interact with atomic nuclei in the material of 
the detector (carbon, lead, iron, water, helium and scintillator). The basic element 
of the detector is a hexagonal plane assembled from parallel triangular scintillating 
strips. The full detector has 200 such planes, and a total weight of 170 tons. The 
detector planes have one of three orientations rotated from each other by 60 degrees, 
and are labeled X, U, or V depending on the orientation. Analyses in the MINERvA 
experiment focus on neutrino interactions in the central tracker, comprising a three- 
ton detector that is fully sensitive to particles produced in the neutrino collisions. 
A charged particle passing through a scintillator produces scintillation light propor- 
tional to its deposited energy. When the light passes through wavelength shifting 
fibers embedded in the scintillator, fluorescent dopants in the fibers emit light at 
a new (green) wavelength, much of which is then transmitted through the optical 
fibers. Optical cables transport that light to photomultiplier tubes located above 
the detector. These signals are used to determine the deposited energy of the par- 
ticle and the position of the particle in two dimensions. In turn, this information 



26,2013 20:58 WSPC/INSTRUCTION FILE Neutrino'to'MPLA 



-1 



is used to reconstruct particle identities and their trajectories in three dimensions 
by combining the X, U and V views to determine the products of the neutrino in- 
teraction. The central tracker is surrounded by calorimeters that have alternating 
layers of metal and scintillator used to contain high energy showering particles and 
measure their energy. The detector was designed to collect more than 16 million 
neutrino events over four years of beam time. 




Fig. 2. The MINERvA detector, (a) Photograph, (b) Cross sectional diagram of the detector 
(c) Events in the detector from a single beam pulse that contains both kinds of events (neutrino 
interactions in the target and in the rock before the target) considered in this measurement. Each 
pixel contains position and energy information, and the display shows just the central region of 
the tracker (only the central portion was used). 



The signal for this measurement is charged-current neutrino interactions that 
contain muons in the final state. The range of these muons is typically many meters 
in the plastic scintillator, so that each muon leaves a signal in at least several tens 
of scintillator strips, and such events are distinguished by their long and straight 
trajectories. The efficiency for reconstructing these long muon tracks is better than 
95%, and there is a very small probability, negligible for the purposes of this demon- 
stration, of finding such a muon candidate with the beam off. 

In this demonstration, most of the signal is from neutrino interactions in the 



26,2013 20:58 WSPC/INSTRUCTION FILE Neutrino'to'MPLA 



DEMONSTRATION OF COMMUNICATION USING NEUTRINOS 5 

rock upstream of MINERvA that produce muons that enter the upstream face 
of the detector. A smaller component of the signal is from neutrino interactions 
that produce muons in the active region of the detector. At the reduced intensity 
of 2.25 x 10 13 protons per pulse used in this study of neutrino communication, 
an average of 0.81 events is registered during each pulse, through the use of a 
software filter to identify these two classes of signal muons that indicate the beam 
is "on". With a time resolution of a few nanoseconds for the muons selected for 
this demonstration of principle, it is possible to disentangle multiple events in the 
8.1 lis beam bursts. An example of multiple events resulting from a single beam 
pulse is shown in Fig. 3. Figure 4 shows a histogram of the number of muons seen 
per beam pulse when the beam is not modulated. The close correspondence between 
data and a Poisson distribution shows the uncorrelated nature of our events, which 
is assumed in the statistical analysis of the communications experiment. 



.-^100-1 








X 

10- 












1- 










ill 


Hit Time (ns) 


Veto 








II ill I III III [I 








80- 



60- 
o 

| 40 
I 20 




















































lii 




















i 





























Fig. 3. Example of track images from the MINERvA detector. The upper panel shows 6 separate 
events occurring at different times during an 8.1 fis neutrino pulse. The time measurements separate 
them easily. A superposition of tracks from these 6 events is shown for the X, U, and V planes as a 
function of module number in the stack in the lower panels. The horizontal axis is distance (about 
3.6 cm/module) along the axis of the central tracker, and the vertical axis is distance (about 1.8 
cm per strip) in directions perpendicular to the central axis. 



The simplest method for encoding information about the neutrino beam is on-off 
keying (OOK). In this scheme, a "1" or a "0" is represented by the presence or ab- 
sence of a beam pulse, respectively. In the NuMI beam line, OOK was implemented 
by controlling proton beam pulses from the Main Injector. Because the beam has 
a specific pulsed time structure, and because there are many independent detector 
segments, the chance of seeing an event that is not caused by the NuMI beam, e.g., 
from cosmic rays, is extremely small. A "1" bit corresponds to a beam pulse with 
an observed event or events, and a "0" bit is a pulse with no events. 



26,2013 20:58 WSPC/INSTRUCTION FILE Neutrino'to'MPLA 



6 

£ 10" 

C 

> 

10 5 
10 4 
I0 1 
10 2 
10 

123456789 10 

# of muons/gate 

Fig. 4. Muons per beam pulse in a run without beam modulation and higher beam intensity than 
available for this experiment. The black histogram is data and the dotted histogram is a Poisson 
distribution for the appropriate mean. The statistical error bars are in most cases smaller than 
the data points. 

If the occurrence of events is modeled as a random Poisson process, this chan- 
nel is mathematically equivalent to the photon counting (or Poisson) channel in 
fTT)l 

communication theory-^ that models direct-detection of optical communications. 
In this model, the probability that no events are observed during a pulse is e _A , 
where A is the expected number of events per pulse when the beam is "on" . Since 
the neutrino detection system produces no errors when a is sent, the probability 
of a bit error, or bit-error rate (BER), is e~ A /2, assuming that 1 and are equally 
likely. 

The structure of the message is shown in Fig. 5a. The 8-character word "neu- 
trino" is expressed in an abbreviated 5-bit code obtained by omitting the first two 
(left-most) bits in the standard 7-bit American Standard Code for Information In- 
terchange (ASCII) code. This 40-bit message is subsequently encoded using the 
convolutional code with rate \ and constraint length 7, that is, in the NASA/ESA 
Planetary Standardp] and expands the message to 92 bits. This encoded message 
is concatenated with a 64-bit pseudo-noise (PN) synchronization sequence to form 
a 156-bit frame. This frame is repeated for the duration of the communication ex- 
periment. The accelerator throughout this study was operated at 25 pulses spaced 
by 2.2 s, followed by a 6.267 s interval to form a 61.267 s "supercycle." 

The data received consists of 3454 records spanning an interval of just over 
142 minutes, along with time stamps for each record. The detector is read out 
whenever a timing gate is received, independent of the presence of any beam. Each 
record contains the number of events observed in the detector over the duration of 
a beam pulse. Decoding the message requires the locations of the frames within the 
data. This is done by searching for the 64-bit PN sequence as a "sync" word that 
signals the end of each frame. If Xq, . . . , xn-i is the entire sequence of transmitted 




26,2013 20:58 WSPC/INSTRUCTION FILE Neutrino'to'MPLA 



DEMONSTRATION OF COMMUNICATION USING NEUTRINOS 7 



156 bit frame 



Coded Message 



Synchronization 
Sequence 



o 



<T3 
1— 
D 



92 bits 



Transmitted Bits 




Raw Received Bits 



64 bits 



Fig. 5. Format and representations of transmitted data, (a) Frame structure of the transmission, 
(b) Depictions of the transmitted and received data. 



bits and Ko, . . . , ifjv-i the corresponding received counts, then K{ should be a 
Poisson random variable with parameter XXi + Ao, where A is the expected number 
of events from a transmitted pulse, and Ao A is the number of background events, 
which is nearly zero in this test. If the counts are conditionally independent, given 
xq, ■ ■ • , Xff— i, the receive r m odel is then mathematically equivalent to the direct- 
detection optical channel ff^l We searched for the end of a known sync word, e.g., 
So, • • • , Sl-i, by finding the maxima of the log-likelihood statistic 

L-l 



'L-l-j 



- r I K. 



(1) 



However, to ensure symbol synchronization throughout the frame we search for two 
sync words spaced exactly by mo = 156 counts, so that the test statistic becomes 
r(m) = A(to) + A (to — too). At the end of a sync word, this statistic has a mean 
fi = LA/2 and variance a 2 = LA/2. To detect a frame, we looked for m such 
that r(m) > /i — 2a. If half the bits are zero and Ao ~ 0, an estimate of A is 
obtained by noting that 1402 events are observed in 3454 counts, so that A « 
2 x 1402/3454 = 0.81. As illustrated in Fig. 6, this statistic identifies 15 frames 
with symbol synchronization, indicated by open circles. There are clearly other 
frames present that might be decoded successfully with more advanced processing; 
nevertheless, we will not pursue this here. 

The 15 synchronized frames are used to recover the message and to estimate the 
properties of the channel. For any nonzero count, the bit is estimated to be 1, and 
otherwise. Figure 5b illustrates message recovery in pictorial form, with each row of 
pixels in the top half of the image representing the transmitted data in one of the 15 
frames, where dark pixels represent ones and light pixels are zeros. The bottom half 
of Fig. 5b shows the corresponding estimated bits at the receiver, where most (78%) 
are received correctly. Errors can be reduced by combining multiple frames and by 



26,2013 20:58 WSPC/INSTRUCTION FILE Neutrino'to'MPLA 




500 1000 1500 2000 2500 3000 3500 
m (count index) 



Fig. 6. Statistic used for synchronization of frames. The open circles identify the locations of 
correctly reconstructed frames. Some frames were not correctly found because of accelerator aborts 
or data acquisition deadtimes. 



using the convolutional error-control code. The optimal way to combine frames is to 
add the corresponding event counts. In fact, the transmitted message is recovered 
perfectly if the results of all the frames are pooled, since events are observed in each 
column where a "1" is transmitted, and no events occur in any column where a "0" 
is transmitted. Consequently, this result demonstrates the ability to communicate 
information using neutrinos. 

Figure 7 illustrates the relationship between the number of frames combined and 
the uncoded bit-error rate (BER) of the channel, which is the average fraction of 
incorrect received bits. Also shown for comparison is the theoretical uncoded BER 



of the Poisson prediction of BER : 



v /2 for A = 0.81. The measured results are in 



close agreement (within expected fluctuations) with the Poisson model. It should be 
noted that 99% of the transmitted bits are decoded correctly using only 5 frames. 
No errors were observed for any combination of 9 or more frames. 

Figure 7 also shows the BER after error-control decoding with the Viterbi 
algorithmic These values are in very good agreement with the simulated Pois- 
son channel for A = 0.81 for the same type of decoding. In this case, «99% of the 
transmitted bits are received correctly after only two frames, and no decoding errors 
are observed when any three frames are combined. 

The theoretical capacity of the photon counting channel with OOK modulation 
and background rate An = is given bjff^ 

(1 - e"*) / A 



C/B = log 2 



1 



exp 



- 1 



(bits/pulse), 



(2) 



where B is the pulse or baud rate (pulses/sec), C is the channel capacity (bits/s), 
and A is the average number of events per pulse defined above. In our experiment 
A w 0.81, which yields a theoretical capacity of C/B w 0.37 bits/pulse. For compar- 



26,2013 20:58 WSPC/INSTRUCTION FILE Neutrino'to'MPLA 



DEMONSTRATION OF COMMUNICATION USING NEUTRINOS 9 




Fig. 7. A comparison of predicted and measured bit-error rates, with and without error-control 
coding. 



ison, the message was decoded with a 1% error rate after two iterations, or for the 
transmission of 2 x 92 = 184 pulses. Since there were only 40 bits of information 
transmitted, the experimental data rate can be estimated as 0.22 bits/pulse, or more 
than half the theoretical capacity of 0.37 bits/pulse. Since the pulse rate was about 
1/2.2 bits per second, the experimental data rate at an error probability of 1% can 
be estimated as 0.22 x (1/2.2) w 0.1 bits per second. We stress that this is only 
a rough estimate, since there are only 105 combinations of 2 frames selected from 
15, and these combinations do not represent independent measurements because of 
reuse of the frames in forming the combinations. 

In general, long-distance communication using neutrinos will favor detectors op- 
timized for identifying interactions in a larger mass of target material than is visible 
to MINERvA and beams that are more intense and with higher energy neutrinos 
than NuMI because the beam becomes narrower and the neutrino interaction rate 
increases with neutrino energy. Of particular interest are the largest detectors, e.g., 
IceCubeJ^that uses the Antarctic icepack to dete ct events, along with muon stor- 
age rings to produce directed neutrino beamsP^^ 

In summary, we have used the Fermilab NuMI neutrino beam, together with 
the MINERvA detector to provide a demonstration of the possibility for digital 
communication using neutrinos. An overall data rate of about 0.1 Hz was realized, 
with an error rate of less than 1% for transmission of neutrinos through a few 
hundred meters of rock. This result illustrates the feasibility, but also shows the 
significant improvements in neutrino beams and detectors required for practical 
applications. 



26,2013 20:58 WSPC/INSTRUCTION FILE Neutrino'to'MPLA 



10 

Acknowledgments 

This work was supported by the Fermi National Accelerator Laboratory, which 
is operated by the Fermi Research Alliance, LLC, under contract No. DE-AC02- 
07CH11359, including the MINERvA construction project, with the United States 
Department of Energy. Construction support also was granted by the United States 
National Science foundation under NSF Award PHY-0619727 and by the University 
of Rochester. Support for participating scientists was provided by NASA, NSF and 
DOE (USA) by CAPES and CNPq (Brazil), by CoNaCyT (Mexico), by CONICYT 
(Chile), by CONCYTEC, DGI-PUCP and LDI-UNI (Peru), by Latin American Cen- 
ter for Physics (CLAF) and by FASI(Russia). Additional support came from Jeffrcss 
Memorial Trust (MK), and Research Corporation (EM). Finally, the authors are 
grateful to the staff of Fermilab for their contribution to this effort, in particular to 
Jim Hylen for his tireless support of the NuMI neutrino beamline. 

References 

1. R. C. Arnold, Science 177, 163 (1972). 

2. A. W. Saenz, H. Uberall, F. J. Kelly, D. W. Padgett, N. Seeman, Science 198, 295 
(1977). 

3. M. Subotowicz, Acta Astronautica 6, 213 (1979). 

4. J. G. Learned, S. Pakvasa, W. A. Simmons, X. Tata, Quarterly Journal of the Royal 
Astro. Soc. 35, 321 (1994). 

5. J. G. Learned, S. Pakvasa, A. Zee, Phys. Lett. B 671, 15 (2009). 

6. P. Huber, Phys. Lett. B 692, 268 (2010). 

7. D. D. Standi, Phys. Rev. D 76, 111701(R) (2007). 

8. J. Jaeckel, J. Redondo, and A. Ringwald, Europhys. Lett. 87, 10010 (2009). 

9. K. Anderson, ct al. The NuML Facility Technical Design Report. FERMILAB-DESIGN- 
1998-01 (1998). 

10. J. R. Pierce, E. C. Posner, and E. R. Rodemich, IEEE Transactions on Information 
Theory IT-27, 61 (1981). 

11. D. J. Costello, Jr., et al. IEEE Transactions on Information Theory IT-44, 2531 
(1998). 

12. G. L. Liu, and H. H. Tan, IEEE Transactions on Communications COM-24, 227 
(1986). 

13. I. Bar-David and G. Kaplan, IEEE Transactions on Information Theory IT-30, 455 
(1984). 

14. F. Halzen and S. R. Klein, Rev. Set. Instrum. 81 081101-24 (2010). 

15. S. Geer, Phys. Rev. D 57, 6989 (1998).